{"paper_id":"8e0007ed-d393-4aae-9110-7030cb8e129f","body_text":"Uterine contractile activity in healthy women throughout the\nmenstrual cycle measured using a novel quantitative two-\ndimensional transvaginal ultrasound speckle tracking method\nCitation for published version (APA):\nRees, C., de Boer, A., Huang, Y., Wessels, B., Blank, C., Kuijsters, N. P. M., Huppelschoten, A. G., Zizolfi, B.,\nForeste, V., di Spiezio Sardo, A., Christophoridis, N., van Vliet, H. A. A. M., Mischi, M., & Schoot, B. C. (2023).\nUterine contractile activity in healthy women throughout the menstrual cycle measured using a novel quantitative\ntwo-dimensional transvaginal ultrasound speckle tracking method. Reproductive BioMedicine Online, 46(1), 115-\n122. https://doi.org/10.1016/j.rbmo.2022.08.104\nDocument license:\nTAVERNE\nDOI:\n10.1016/j.rbmo.2022.08.104\nDocument status and date:\nPublished: 01/01/2023\nDocument Version:\nPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)\nPlease check the document version of this publication:\n• A submitted manuscript is the version of the article upon submission and before peer-review. There can be\nimportant differences between the submitted version and the official published version of record. People\ninterested in the research are advised to contact the author for the final version of the publication, or visit the\nDOI to the publisher's website.\n• The final author version and the galley proof are versions of the publication after peer review.\n• The final published version features the final layout of the paper including the volume, issue and page\nnumbers.\nLink to publication\nGeneral rights\nCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners\nand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.\n            • Users may download and print one copy of any publication from the public portal for the purpose of private study or research.\n            • You may not further distribute the material or use it for any profit-making activity or commercial gain\n            • You may freely distribute the URL identifying the publication in the public portal.\nIf the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please\nfollow below link for the End User Agreement:\nwww.tue.nl/taverne\nTake down policy\nIf you believe that this document breaches copyright please contact us at:\nopenaccess@tue.nl\nproviding details and we will investigate your claim.\nDownload date: 14. Jun. 2026\n\n115  RBMO  VOLUME 46  ISSUE 1  2023\n1 Department of Obstetrics and Gynaecology, Catharina Hospital, Eindhoven, the Netherlands\n2 Laboratory of Biomedical Diagnostics, Department of Electrical Engineering, Eindhoven University of Technology, \nEindhoven, the Netherlands\n3 Department of Reproductive Medicine, Ghent University Hospital, Ghent, Belgium\n4 Department of Public Health – School of Medicine, University Federico II, Naples, Italy\n5 Embryolab Fertility Center Thessaloniki, Greece\n© 2022 Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.\n*Corresponding author. E-mail address: connie.rees@catharinaziekenhuis.nl (C. O. Rees). https://doi.org/10.1016/j.\nrbmo.2022.08.104 1472-6483/© 2022 Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.\nDeclaration: The authors report no financial or commercial conflicts of interest.\nKEYWORDS\nMenstrual cycle\nSpeckle tracking\nTransvaginal ultrasound\nTVUS\nUterine contractile function\nUterine peristalsis\nARTICLE\nUterine contractile activity in healthy women \nthroughout the menstrual cycle measured \nusing a novel quantitative two-dimensional \ntransvaginal ultrasound speckle tracking method\nBIOGRAPHY\nConnie Rees received her degree in Medicine from Utrecht University. Subsequently, she \nstarted her PhD on adenomyosis and uterine contractility at Catharina Hospital Eindhoven, \nthe Netherlands. The research is led by Professor Dick Schoot, Professor Massimo Mischi \nand Professor Huib van Vliet from Eindhoven University of Technology and Ghent University. \nConnie Odette Rees1,2,3,*, Anna de Boer1, Yizhou Huang2, Blijke Wessels2, \nCeline Blank2,3, Nienke Kuijsters2, Aleida Huppelschoten1, Brunella Zizolfi4, \nVirginia Foreste4, Attilio Di Spiezio Sardo4, Nikos Christoforidis5, \nHubertus van Vliet1,3, Massimo Mischi2, Benedictus Christiaan Schoot1,2,3\nKEY MESSAGE\nUterine contractility measured by objective transvaginal ultrasound speckle tracking included the novel features coordination, \ndirection and velocity. Menstrual cycle uterine contractility was highest in the periovulatory phase, and lowest in the late luteal \nphase. Future studies could investigate uterine contractility by this method in women with infertility or abnormal uteri.\nABSTRACT\nResearch question: To explore normal uterine contractile function across the menstrual cycle using a novel quantitative ultrasound method.\nDesign: This multicentre prospective observational study took place in three European centres from 2014 to 2022. Uterine \ncontraction frequency (contractions/minute), amplitude, direction (cervix-to-fundus, C2F; fundus-to-cervix; F2C), velocity and \ncoordination were investigated. Features were extracted from transvaginal ultrasound recordings (TVUS) using speckle tracking. \nPremenopausal women ≥18 years of age, with normal, natural menstrual cycles were included. A normal cycle was defined as: \nregular (duration 28 ± 2 days), no dysmenorrhoea, no menometrorrhagia. Four-minute TVUS were performed during the menstrual \nphase, mid-follicular, late follicular phase, early luteal phase and/or late luteal phase. Of the 96 recordings available from 64 \nwomen, 70 were suitable for inclusion in the analysis.\nResults: Contraction frequency (for the posterior wall) and velocity (for the anterior uterine wall in the F2C direction) were highest \nin the late follicular phase and lowest in the menstrual and late luteal phases (1.61 versus 1.31 and 1.35 contractions/min, P < 0.001 \nand 0.81 versus 0.67 and 0.62 mm/s, P < 0.001, respectively). No significant difference was found for contraction amplitude. \nContraction coordination (simultaneous contraction of the anterior and posterior walls in the same direction) was least coordinated \nin the mid-follicular phase (P = 0.002).\nConclusions: This is the first study to objectively measure uterine contraction features in healthy women during the natural menstrual cycle \non TVUS. Likewise, it introduces contraction coordination as a specific feature of uterine peristalsis. Differences in uterine contractility \nacross the menstrual cycle are confirmed, with highest activity seen in the late follicular phase, and lowest in the late luteal phase.\n\n\n116  RBMO  VOLUME 46  ISSUE 1  2023\nINTRODUCTION\nI\nn a healthy uterus, rhythmic \ncontractions change in rhythm and \nintensity during the menstrual cycle \nto support sperm propagation and \nembryo implantation (Bulletti and De \nZiegler, 2006; Fanchin and Ayoubi, \n2009; Kuijsters et al., 2017; van Gestel \net al., 2003). However, no study thus \nfar has been able to comprehensively \ncharacterize all aspects of uterine \ncontractions during the menstrual \ncycle; therefore, these characteristics \nremain largely speculative, based on \nheterogenous studies. Furthermore, \nresearch into their characteristics have \nbeen hampered by the subjectivity of the \navailable measurement tools (Kuijsters \net al., 2017).\nIt has been suggested that aberrant \nuterine peristalsis or ‘dysperistalsis’ is \nassociated with reduced fertility and/\nor symptoms such as dysmenorrhoea \n(Kissler et al., 2007). Up to now, there \nexists no quantifiable marker for \ndysperistalsis, which is variously defined \nby previous investigators.\nMultiple methods have been used to \nvisualize and assess uterine contractions \nand their different characteristics, one of \nwhich is transvaginal ultrasound (TVUS). \nA recent study, however, showed that \nmedical professionals shared only mild \nagreement on the direction and timing \nof uterine peristalsis by subjective visual \ninspection of TVUS recordings (Kuijsters \net al., 2020). Although visual inspection \nof TVUS can provide a number of \nperistalsis parameters (frequency and \ndirection), it is generally qualitative \nand unsuitable to quantify contraction \namplitude or velocity. Furthermore, \ncontraction coordination – the \nsynchronized movement of the anterior \nand posterior walls of the uterus – has \nnever been investigated. There is thus \na need for an objective, quantifiable \nmethod of uterine contraction \nassessment, preferably in a non-invasive \noperator- and patient-friendly way.\nRecently published data by the current \nauthor group presented a novel method \nfor assessing uterine contractility, using \ntwo-dimensional (2D) TVUS and speckle \ntracking techniques (Huang et al., 2018b; \nSammali et al., 2018a, 2018b, 2019). \nThis has been tested and (externally) \nvalidated in IVF patients prior to embryo \ntransfer (Blank et al., 2020; Sammali \net al., 2021). This method is able to \nquantitatively assess features such as \ncontraction frequency and amplitude, \nin addition to a novel set of features: \ncontraction coordination, direction and \nvelocity. Coordination is defined as the \nsynchronized simultaneous movement \nof the anterior and posterior uterine \nwalls, where the value reflects the degree \nof synchronicity of coordination. This \naspect of uterine movement is potentially \nof clinical relevance for the assessment of \n(normal) uterine function.\nDifferences in uterine contractility have \nbeen shown to have a strong association \nwith ongoing pregnancy in an IVF \npopulation (Blank et al., 2020). This study \nexplores the characteristics of normal \nuterine contractile function across the \nmenstrual cycle in healthy, nulliparous \nwomen using this quantitative method, \nwith a focus on the novel feature of \ncoordination as a possible measure of \ndysperistalsis.\nMATERIALS AND METHODS\nStudy objectives\nTo evaluate uterine contraction features \n(frequency, amplitude, velocity, direction \nand coordination) using a dedicated \nspeckle tracking algorithm by 2D \nTVUS measurement in healthy women \nthroughout the menstrual cycle.\nStudy design and setting\nMulticentre observational prospective \ncohort study carried out in the \noutpatient gynaecology departments of \nthe Catharina Hospital in Eindhoven, \nthe Netherlands; the University of \nNaples, Federico II Naples, Italy; and \nthe Embryolab Fertility Centre in \nThessaloniki, Greece.\nParticipants\nBetween September 2014 and \nJanuary 2022, 74 healthy women were \nincluded from the gynaecological \noutpatient departments of the \nparticipating centres. Women were \nincluded if they were ≥18 years of age, \npremenopausal and had a normal, \nnatural menstrual cycle. A normal \ncycle was defined as: regular (duration \n28 ± 2 days), no dysmenorrhoea and \nno menometrorrhagia. Exclusion criteria \nwere: (i) pregnancy, (ii) diagnosed \nwith a mental disorder, (iii) significant \nlanguage barrier, (iv) use of oral hormonal \ncontraceptives or intrauterine device, \n(v) use of other (hormonal) medication \naffecting the uterus, or (vi) uterine \npathology (congenital or otherwise, \ne.g. leiomyomas, adenomyosis), based \non morphological uterine sonographic \nassessment (MUSA) (Van den Bosch \net al., 2019) criteria. Ultrasound scans of \nthe included women were also assessed \nretrospectively by experts to confirm the \nabsence of uterine abnormalities.\nSeventy-four women were enrolled \nin the study, of which 64 ultimately \nunderwent TVUS recording at different \nphases of the menstrual cycle. This \nresulted in a total of 96 completed TVUS \nrecordings across cycle phases. Eighteen \nrecordings were subsequently excluded \ndue to insufficient ultrasound quality \nfor the analysis. Reasons for exclusion \ndue to recording quality included: \nshadow across the endometrial lining, \nout-of-plane motion or insufficient \nresolution of the images. Additionally, \neight recordings were excluded due to \nsuspected uterine abnormalities or use \nof hormonal contraceptive methods. \nFIGURE 1 presents a flow diagram of \npatient inclusion. Overall, 70 out of 96 \nconducted recordings from 64 women \nwere included in the analysis.\nData sources and measurements\nTVUS measurement\nTVUS was performed during several \nphases of the menstrual cycle: the \nmenstrual phase (cycle days 1–5), mid-\nfollicular phase (cycle days 6–10), late \nfollicular phase (cycle days 11–14), early \nluteal phase (cycle days 15–20) and late \nluteal phase (cycle days 21–28). During \neach session, 4-min video recordings \nof the uterus in the mid-sagittal section \nwere made. The ultrasound machines \nused were an Accuvix WS80A with \nElite for Women's Health (Samsung \nMedison, Seoul, Korea) equipped with \na V5-9 transvaginal probe (bandwidth \n5–9 MHz) or a Voluson S10™ Expert \n(GE Healthcare, Zipf, Austria) equipped \nwith a RIC5-9W-RS probe (bandwidth \n3.8–9 .3 MHz).\nFeature extraction\nVarious uterine contractility features were \nextracted from the gathered ultrasound \nrecordings using a quantitative dedicated \nspeckle tracking algorithm previously \ndeveloped and implemented in Matlab \nsoftware (Mathworks, Natick, USA). \nThe full details of the methodology of \nfeature extraction have been described \nin detail in previously published works \n\n RBMO  VOLUME 46  ISSUE 1  2023  117\n(Blank et al., 2020; Huang et al., 2018a, \n2022; Sammali et al., 2018a, 2019a, \n2019b). Simply put, speckle tracking \nmeasures the displacement of image \n‘speckles’ (such as those seen in various \nshades of grey on ultrasound images) \nover time. Speckle movement reflects \nmovement of the imaged tissue, which \nin this case is movement of the uterine \nmyometrium.\nFor each ultrasound recording, a grid of \ntracking markers was manually positioned \nover the uterine junctional zone along \nthe endometrium, known to be the most \ncontractile part of the uterus (see FIGURE 2 \nfor an illustrative example). Grid markers \nwere placed 5 mm from the fundus along \nthe endometrial border. The grid markers \nwere coupled in pairs, and distance and \nstrain signals were derived between each \npair in both the longitudinal and radial \ndirections (FIGURE 3). Several contraction \nfeatures were extracted from the \nmeasured strain signals as described \nbelow. Previous analyses of inter- and \nintra-observer variability in the placement \nof the grid markers showed a high level \nof correlation, making the method both \nreproducible and reliable (Huang et al., \n2022) (see Appendix A).\nEuclidean distance was used to derive \nthe distance between each pair, resulting \nin absolute motion estimates. The strain, \n∈, was defined as the relative variation \nof the distance, d, between the tracked \nblocks as: ∈ = d(j)−d(j−1)/d(j−1), where \nd(j) and d(j–1) indicate the distance \nbetween the tracked blocks at the \ncurrent frame (j) and previous frame (j–1), \nrespectively.\nContraction frequency\nFrequency features were analysed \nseparately for the anterior and posterior \nFIGURE 1 Flow chart of patient recruitment and inclusion. TVUS = transvaginal ultrasound.\nFIGURE 2 Ultrasound image of the uterus in the mid-sagittal section. Placement of the speckle \ntracking grid (red dots) along the endometrial border (blue line) at 5 mm (green line) from the \napex of the fundus.\nFIGURE 3 Ultrasound images of the uterus in the mid-sagittal section with depiction of contractions in the longitudinal (left) and radial (right) \ndirection.\n\n118  RBMO  VOLUME 46  ISSUE 1  2023\nwalls of the uterus, in the longitudinal and \nradial directions. Only the longitudinal \ndirection of contractions is presented \nhere. Frequency-related features are \nreported as contractions per minute \n(Sammali et al., 2019b). Further technical \ndetails about feature extraction, as well as \npre-processing analysis, can be found in \nSammali et al. (2018a).\nContraction amplitude\nContraction amplitude features \nreflect the relative strength of uterine \ncontraction. Amplitude of contraction \nwas assessed by calculating the SD of the \nstrain signal in the longitudinal and radial \ndirections from its frequency spectrum \n(Parseval's theorem) (Hu and Fan, 2010). \nA higher value reflected stronger uterine \ncontractions. Results are reported for \ncontractions in the longitudinal direction, \nseparately for the anterior and posterior \nuterine walls.\nContraction direction\nUterine peristalsis is thought to \npropagate mainly in one of two \ndirections: either fundus-to-cervix \n(F2C) or cervix-to-fundus (C2F). The \ncontraction direction was estimated \nby analysis of the radial strain signal \nrepresentation in the spatiotemporal \nfrequency domain, where the spatial \ndomain is intended along the longitudinal \ndirection of the uterus (Huang et al., \n2022). The ratio between the strain signal \nenergy in the quadrants corresponding \nto the two propagation directions (C2F \nand F2C) provided a global measure \nof the dominant propagation direction \nin each wall (posterior and anterior) \nseparately (Huang et al., 2022). Basically, \na more positive value represented \nmovement predominantly in the F2C \ndirection, whereas a more negative value \nrepresented movement predominantly \nin the C2F direction. A value around \nzero represented movement that did \nnot show a predominant direction, being \neither a circular movement, or standing \nor opposing contractions.\nContraction velocity\nVelocity, the propagation speed of the \nperistaltic waves in a certain direction \n(C2F or F2C, in mm/s), was calculated for \nmovement in the anterior and posterior \nwalls. This was again done by analysing \nthe radial strain signal representation in \nthe spatiotemporal frequency domain, \nwhere the spatial domain is intended \nalong the longitudinal direction of the \nuterus (Sammali et al., 2018a). The \nanalysis was performed over a window \nof 20 s sliding over the full recording \ntime. Subsequently, the median velocities \nin the C2F and F2C directions were \ncalculated by averaging velocities over \ntime in the corresponding directions; a \nhigh value reflected increased velocity \nin the reported direction. Results were \nreported separately for the anterior and \nposterior uterine walls.\nContraction coordination\nIn addition to the features described \nabove, the aim was to also assess the \ncoordination of uterine contraction. \nThis is the first attempt at a quantifiable \nmeasurement of coordination of uterine \nmovement to date. In order to quantify \nthis, an assessment was done of whether \nthe anterior and posterior walls of the \nuterus were moving synchronously or \nasynchronously. This was accomplished \nby estimating the time evolution of the \nestimated propagation direction over \nthe anterior and posterior walls using \na running window of 20 s. The two \nresulting evolutions were then compared \nusing a similarity measure. This resulted \nin a feature defining the uterine \ncontraction coordination depending \non the adopted similarity measure: the \nmean square error (MSE). Two additional \ncoordination features (Hausdorff distance \nmetric and cross-correlation) were also \ninvestigated, and are shown in Appendix \nA. Again, full details of the technical \nbackground of these units have been \npublished elsewhere (Huang et al., 2022). \nA higher value reflected decreased \ncontraction coordination.\nStudy outcomes\nThe primary outcomes investigated \nwere the following uterine contraction \nfeatures, compared between the four \nmenstrual phases: (i) frequency, in \ncontractions/minute; (ii) amplitude \n(unitless); (iii) direction (unitless, \nwhereby >0.0 globally represents F2C \nmovement, and <0.0 represents C2F \nmovement); (iv) median velocity (mm/s); \n(v) coordination, in MSE.\nStatistical methods\nStatistical analysis was performed using \nSPSS Statistics for Windows, Version \n27 (IBM Corp., Armonk, NY, USA). The \nShapiro–Wilk test was first employed to \ntest the normality of the distributions. \nComparison of the outcome measures \n(frequency, amplitude, direction, \ncoordination and velocity) between \nthe various phases was done using \nthe Kruskal–Wallis test if abnormally \ndistributed, and a one-way analysis of \nvariance (ANOVA) if normally distributed \n(with Bonferroni correction). Statistical \nsignificance was defined as a P-value \n<0.05. This study is reported according \nto the STROBE (Strengthening the \nReporting of Observational Studies in \nEpidemiology) guidelines (Von Elm et al., \n2007).\nEthical approval\nThis study received ethical approval from \nthe local and regional ethical committees \nof participating centres, with study \nnumber NL52466.100.15 on 15 July 2020 \n(the Netherlands), 12 May 2021 (Greece) \nand September 9th, 2021 (Italy). All \nparticipants gave informed consent prior \nto study participation.\nRESULTS\nPatient characteristics and \nrecruitment\nUltimately (see FIGURE 1), 70 recordings \nfrom 64 women were available for \nanalysis. TABLE 1 presents an overview of \nthe characteristics of these women.\nUterine contraction features\nTABLES 2–6 present an overview of the \nvalues found per contraction feature \nacross the menstrual cycle phases, for \nthe features of frequency, direction, \nvelocity, amplitude and coordination.\nContraction frequency, velocity and \ncoordination differed significantly \nbetween menstrual phases. No significant \ndifferences were found between cycle \nphases for amplitude or direction.\nContraction frequency\nThe overall values of contraction \nfrequency per menstrual phase can \nbe seen in TABLE 2. The highest mean \ncontraction frequency (1.61, SD 0.17, \nP < 0.001) was found in the late follicular \nphase in the posterior wall. The phase \nwith the lowest mean contraction \nfrequency was the late luteal phase (1.28, \nSD 0.13) in the anterior wall, P = 0.003). \nThe early luteal and menstrual phases \nhad comparable contraction frequencies \n(P > 0.05).\nContraction amplitude\nThe overall values of contraction \namplitude per cycle phase can be seen \nin TABLE 3. No significant differences were \nfound between cycle phases. The mean \ncontraction amplitude was 0.044–0.062 \n\n RBMO  VOLUME 46  ISSUE 1  2023  119\n(SD 0.011–0.016) in the late follicular \nphase and 0.036–0.062 (SD 0.013–0.024) \nin the late luteal phase (P > 0.05).\nContraction direction and velocity\nContraction direction did not seem to \ndiffer significantly between menstrual \nphases (TABLE 4, P > 0.05). During \nthe menstrual phase, direction of \ncontraction showed a trend towards F2C \ncontractions. In other phases mainly C2F \ncontractions were seen. Contraction \nvelocity overall differed significantly \nacross cycle phases. The velocity of \ncontractions was significantly higher in \nthe late follicular phase in all directions \n(see TABLE 5, P < 0.001, P = 0.021, 0.004 \nand 0.026, respectively), and lowest in \nthe late luteal phase.\nContraction coordination\nThe contraction coordination values are \nshown for all cycle phases in TABLE 6. MSE \nshowed a significant difference across \nthe cycle phases, with a significantly \n(P = 0.011) reduced coordination of \ncontractions during the late follicular \nphase compared with the menstrual and \nlate luteal phases. Further coordination \nparameters did not differ significantly \nbetween cycle phases (see Appendix B).\nDISCUSSION\nThe results of this exploratory study \nsuggest a preliminary range of normal \nreference values in a healthy population of \nwomen without hormonal contraception \nTABLE 1 PATIENT CHARACTERISTICS FOR ANALYSED PATIENTS ( N = 64)\nCharacteristic\nPatients per participating centre\n Netherlands 33 (51.6)\n Italy 24 (37 .5)\n Greece 7 (10.9)\n Age (years) 34.1 (6.3)\n BMI (median, IQR) 23.0 (3.75)\nParity\n Nulliparous 34 (53.1)\n Multiparous 9 (14.1)\n Missing 11 (17 .2)\n Cycle duration (days) 28.1 (1.7)\n Cycle day menses measurement 2.42 (1.13)\n Cycle day mid-follicular measurement 8.33 (0.94)\n Cycle day late follicular measurement 12.31 (0.85)\n Cycle day early luteal measurement 16.33 (3.95)\n Cycle day late luteal measurement 27 .50 (1.00)\nUterine measurements\n Uterine length (mm) 71.44 (10.77)\n Uterine height (mm) 35.96 (6.18)\n Uterine width (mm) 63.10 (1.06)\nEndometrial thickness (mm) (median, IQR) per cycle phase Menses: 2.00 (0)\nMid-follicular: 5.34 (1.52)\nLate follicular: 7 .46 (2.77)\nEarly luteal: 10.02 (3.92)\nLate luteal: 6.60 (2.44)\nData are presented as n (%) or mean (SD) unless otherwise stated.\nBMI = body mass index; IQR = interquartile range.\nTABLE 2 CONTRACTION FREQUENCY ACCORDING TO MENSTRUAL CYCLE PHASE\nMenstrual \n(n = 4)\nMid-follicular \n(n = 11)\nLate follicular \n(n = 26)\nEarly luteal \n(n = 14)\nLate luteal \n(n = 15)\nP-value (one-way \nANOVA)\nContraction frequency, longitudinal, \nanterior wall (contractions/min)\n1.31 (0.08) 1.46 (0.12)a 1.46 (0.14)a 1.40 (0.14) 1.28 (0.13) 0.003\nContraction frequency, longitudinal, \nposterior wall (contractions/min)\n1.31 (0.13) 1.54 (0.14)a,b 1.61 (0.17)a,b 1.45 (0.17) 1.35 (0.19) <0.001\nData are presented as mean (SD).\na Significant difference versus late luteal phase.\nb Significant difference versus menstrual phase.\nTABLE 3 CONTRACTION AMPLITUDE ACCORDING TO MENSTRUAL CYCLE PHASE\nMenstrual \n(n = 4)\nMid-follicular \n(n = 11)\nLate follicular \n(n = 26)\nEarly luteal \n(n = 14)\nLate luteal \n(n = 15)\nP-value (one-way \nANOVA)\nSD in strain longitudinal direction anterior 0.050 (0.015) 0.049 (0.010) 0.062 (0.016) 0.056 (0.013) 0.062 (0.024) 0.141\nSD in strain longitudinal direction posterior 0.042 (0.006) 0.036 (0.010) 0.043 (0.012) 0.041 (0.014) 0.040 (0.014) 0.240\nSD in strain radial direction anterior 0.041 (0.014) 0.038 (0.010) 0.045 (0.012) 0.047 (0.021) 0.038 (0.013) 0.266\nSD in strain radial direction posterior 0.041 (0.011) 0.037 (0.010) 0.044 (0.011) 0.044 (0.023) 0.036 (0.013) 0.218\nData are presented as mean (SD).\n\n120  RBMO  VOLUME 46  ISSUE 1  2023\nTABLE 6 CONTRACTION COORDINATION ACCORDING TO MENSTRUAL CYCLE PHASE\nMenstrual \n(n = 4)\nMid-follicular \n(n = 11)\nLate follicular \n(n = 26)\nEarly luteal \n(n = 14)\nLate luteal \n(n = 15)\nP-value (one-way \nANOVA)\nMean square error 0.15 (0.04)b 0.28 (0.75)c 0.24 (0.12)a,c 0.20 (0.08) 0.18 (0.07) 0.011\nData are presented as mean (SD).\na Significant difference versus late luteal phase.\nb Significant difference versus late follicular phase.\nc Significant difference versus menstrual phase.\nand normal uteri. A novel, reproducible \nand objective method based on \nultrasound speckle tracking is presented. \nIt is possible to characterize uterine \ncontraction amplitude and frequency, \nas well as coordination, direction \nand velocity. Coordination, direction \nand velocity of uterine contractions \nare features that have never before \nbeen quantified in this context. The \ncurrent results show that contraction \nfrequency and velocity are highest in \nthe late follicular phase and lowest in \nthe menstrual and late luteal phases. \nCoordination seems to be negatively \naffected by contractions with higher \nfrequency and velocity in the late follicular \nphase compared with other phases. \nAmplitude and contraction direction in \nthis population do not show significant \nvariations across the menstrual cycle.\nThe findings are generally in accordance \nwith the existing literature concerning \nuterine contractile activity in the healthy \nuterus. Previously described methods \nto assess uterine contractility have \nassessed some subsets of the features \npresented here (Blank et al., 2020); \nhowever, this is the first study where all \nthe presented features are quantified \nand evaluated (Kuijsters et al., 2017). The \nnovel features for the characterization of \ndifferent uterine activity and associated \npatterns – coordination, direction and \nvelocity – could form a new avenue for \nresearch and knowledge into uterine \nfunction. The TVUS method presented \nfor the quantitative analysis of uterine \ncontractions is also easily reproducible \n(Huang et al., 2022): quick, objective and \npatient-friendly. It is potentially possible \nto integrate into routine gynaecological \npractice (after sufficient training), and \ndoes not require extensive skill or \nexpertise.\nThe main limitation of the results \npresented here is the small sample size \nof the study population. However, it is \nbelieved that the results presented are \nvalid due to their general accordance with \nthe currently accepted patterns of uterine \nperistalsis throughout the menstrual \ncycle. In addition, most patients received \nan ultrasound in only one phase of the \nmenstrual cycle for a 4-minute time \nframe, and therefore it was not possible \nto conduct a within-subjects comparison. \nIt could be debated how far this relatively \nshort recording is representative of the \nbehaviour of the uterus during this phase \nin general, however a subanalysis with \nrepeated recordings within subjects was \nconducted in previous work (Huang \net al., 2022), with comparable results. \nAdditionally, the majority of ultrasounds \nwere conducted in the late follicular \nphase, which may affect the significance \nof results. Furthermore, due to the \nnovel and still experimental nature of \nthe quantitative analysis employed in this \nstudy, its clinical application in routine \npractice is not yet possible. It was also \nnecessary to exclude a significant number \nof recordings from analysis (n = 18) due \nto insufficient quality of the ultrasounds, \nwhich indicates that there is a learning \nTABLE 4 CONTRACTION DIRECTION ACCORDING TO MENSTRUAL CYCLE PHASE\nMenstrual \n(n = 4)\nMid-follicular \n(n = 11)\nLate follicular \n(n = 26)\nEarly luteal \n(n = 14)\nLate luteal \n(n = 15)\nP-value (one-way \nANOVA)\nDirection anterior walla 0.085 (0.288) –0.100 (0.379) –0.032 (0.396) –0.054 (0.297) 0.084 (0.153) 0.669\nDirection posterior walla 0.013 (0.279) –0.270 (0.252) –0.207 (0.180) –0.206 (0.179) –0.061 (0.242) 0.300\nPredominant direction F2C C2F C2F C2F None n/a\nData are presented as mean (SD).\na A value under 0.0 reflects movements predominantly in the cervix-to-fundus (C2F) direction, whereas a value higher than 0.0 reflects movement predominantly in the \nfundus-to-cervix (F2C) direction. Values between –0.1 and 0.1 reflect no predominant direction, or standing/opposing contractions.\nTABLE 5 CONTRACTION VELOCITY ACCORDING TO MENSTRUAL CYCLE PHASE\nMenstrual \n(n = 4)\nMid-follicular \n(n = 11)\nLate follicular \n(n = 26)\nEarly luteal \n(n = 14)\nLate luteal \n(n = 15)\nP-value \n(Kruskal–Wallis)\nFundus-to-cervix propagation anterior (mm/s) 0.67 (0.10) 0.77 (0.26) 0.81 (0.31)a 0.71 (0.17) 0.62 (0.08) <0.001\nFundus-to-cervix propagation posterior (mm/s) 0.69 (0.09) 0.73 (0.23) 0.85 (0.21) 0.73 (0.12) 0.67 (0.16) 0.021\nCervix-to-fundus propagation anterior (mm/s) 0.68 (0.06) 0.80 (0.30) 0.82 (0.29)a 0.71 (0.14) 0.65 (0.13) 0.004\nCervix-to-fundus propagation posterior (mm/s) 0.65 (0.11) 0.78 (0.31) 0.86 (0.33)a 0.74 (0.11) 0.66 (0.15) 0.026\nData are presented as median (IQR).\na Significant difference versus late luteal phase.\n\n RBMO  VOLUME 46  ISSUE 1  2023  121\ncurve which could (initially) affect \nclinical useability. In some cases this was \navoidable (e.g. insufficient resolution or \nout-of-plane motion), but incidentally it \nis not possible to analyse contractions \ndespite good ultrasound technique (for \ninstance due to the orientation of the \nuterus, or shadows caused by intestinal \ncontents, for example). It is also not \nyet feasible to gain contraction feature \nresults in real time while performing the \nultrasound scan, as the implemented \nanalysis still relies on offline, post-\nultrasound data processing. In the future, \nsteps need to be taken to make this TVUS \nspeckle tracking method for quantitative \nanalysis of uterine contractions usable in \ndaily clinical practice.\nThe results presented here provide \nfurther insight into uterine behaviour \nat different phases of the menstrual \ncycle, whereby each cycle phase shows \nan individual contraction pattern. The \nresults clearly show that the late follicular \nphase is the most active, with the highest \ncontraction frequency and velocity. One \ncould surmise that these features are \nthus important for the sperm transport \nand ovulation that occurs in this period \nof the menstrual cycle. Furthermore, the \nrelatively reduced activity in the late luteal \nto menstrual phases suggests a relevance \nof these characteristics with regard \nto facilitation of embryo implantation \nand/or menstruation symptoms. The \ncoordination feature has not been \ninvestigated before; however, these initial \nresults show that increased contraction \nfrequency and velocity seem to be \naccompanied by reduced coordination of \ncontractions. The clinical importance of \nsimultaneous (coordinated) anterior and \nposterior contractions and how this could \nrelate to fertility outcomes or clinical \nsymptoms, for example, merits further \ninvestigation. Potentially, this coordination \nfeature could be seen as a measure of \nuterine dysperistalsis, which has been \npreviously described in patients with \ninfertility and endometriosis (Kissler et al., \n2006, 2007; Leyendecker et al., 1996) \nwith significant clinical consequences, \nespecially with regards to fertility.\nNow that it is possible to suggest \npreliminary reference values for uterine \nperistalsis in a normal menstrual cycle, \nit is possible to better assess how and if \nuterine contractile activity is abnormal in \ndifferent populations. Previous work by \nthis research group has assessed uterine \nperistalsis in IVF patients, which showed \npromise with regards to prediction of IVF \ntreatment success (Blank et al., 2020). \nFuture works will be able to compare \nhow uterine contractions differ between \nfertile and infertile populations, also \nrelative to the preliminary reference \nvalues in a normal menstrual cycle \npresented here, potentially identifying \ntreatment target points, and perhaps \nuncovering a new facet of infertility \naetiological mechanisms.\nAlthough this study focused on \nhealthy women with normal uteri, \nuterine peristalsis assessment and \ncharacterization can also represent a \nvaluable diagnostic tool in the context \nof common pathological conditions \nof the uterus, such as adenomyosis, \nendometriosis or uterine fibroids. The \neffect of these conditions on uterine \nfunction (and disease symptoms such \nas dysmenorrhoea and infertility) may \nin fact be reflected in altered uterine \nperistalsis patterns, such as coordination. \nDedicated clinical trials can be designed \nto investigate the potential of the \nproposed features for the diagnosis \nof uterine diseases and dysfunctions. \nDifferences in uterine contractions \ncould be an explanatory factor for the \nsymptomatology in certain uterine \ndisorders, and thereby become a target \nfor patient-tailored treatment.\nIn summary, preliminary reference values \nof uterine contraction features in healthy \nwomen during the natural menstrual \ncycle are suggested. The current study \nserves as a standard to which uterine \nperistalsis in infertile women or women \nwith abnormal uteri can be compared, \npotentially identifying treatment targets \nand aetiological mechanisms yet \nunexplored. Furthermore, the study \npresents novel uterine contraction \nfeatures that can be used to assess the \npresence (or absence) of normal uterine \ncontractility, namely coordination, \ndirection and velocity.\nACKNOWLEDGEMENTS\nThe authors received an unrestricted \ngrant from GE Healthcare Austria.\nSUPPLEMENTARY MATERIALS\nSupplementary material associated \nwith this article can be found, in \nthe online version, at doi:10.1016/j.\nrbmo.2022.08.104.\nREFERENCES\nBlank, C., Sammali, F., Kuijsters, N., Huang, Y., \nRabotti, C., de Sutter, P ., Mischi, M., Schoot, B. \nAssessment of uterine activity during IVF by \nquantitative ultrasound imaging: a pilot study. \nReprod. Biomed. Online 2020; 41: 1045–1053. \ndoi:10.1016/J.RBMO.2020.08.006\nBulletti, C., De Ziegler, D. Uterine contractility \nand embryo implantation. Curr. Opin. Obstet. \nGynecol. 2006; 18: 473–484. doi:10.1097/01.\ngco.0000233947 .97543.c4\nFanchin, R., Ayoubi, J.M. Uterine dynamics: \nimpact on the human reproduction process. \nReprod. Biomed. Online 2009; 18: 57–62. \ndoi:10.1016/S1472-6483(10)60450-6\nHu, L.Y., Fan, H.Y., 2010. Inversion formula and \nParseval theorem for complex continuous \nwavelet transforms studied by entangled state \nrepresentation. Chinese Phys. B 19, 074205. \nhttps://doi.org/10.1088/1674-1056/19/7/074205\nHuang, Y., Rees, C., Sammali, F., Blank, C., \nSchoot, D., Mischi, M. Characterization of \nuterine peristaltic waves by ultrasound strain \nanalysis. IEEE Trans. Ultrason. Ferroelectr. \nFreq. Control. 2022; 69: 2050–2060. \ndoi:10.1109/TUFFC.2022.3165688\nHuang, Y., Sammali, F., Blank, C., Kuijsters, \nN., Rabotti, C., Schoot, B.C., Mischi, M. \nQuantitative ultrasound imaging and \ncharacterization of uterine peristaltic waves.  \nIEEE International Ultrasonics Symposium, IUS \n2018: 1–4. doi:10.1109/ULTSYM.2018.8580218\nHuang, Y., Sammali, F., Kuijsters, N.P .M., Blank, \nC., Schoot, B.C., Mischi, M. Quantitative \nmotion analysis of the uterus by optical flow \nand two-dimensional strain mapping.  2018 \nIEEE International Symposium on Medical \nMeasurements and Applications (MeMeA) \n2018: 1–5. doi:10.1109/MeMeA.2018.8438735\nKissler, S., Hamscho, N., Zangos, S., Wiegratz, \nI., Schlichter, S., Menzel, C., Doebert, N., \nGruenwald, F., Vogl, T.J., Gaetje, R., Rody, A., \nSiebzehnruebl, E., Kunz, G., Leyendecker, G., \nKaufmann, M. Uterotubal transport disorder \nin adenomyosis and endometriosis – a cause \nfor infertility. BJOG 2006; 113: 902–908. \ndoi:10.1111/j.1471-0528.2006.00970.x\nKissler, S., Zangos, S., Wiegratz, I., Kohl, J., Rody, \nA., Gaetje, R., Doebert, N., Wildt, L., Kunz, \nG., Leyendecker, G., Kaufmann, M. Utero-\ntubal sperm transport and its impairment \nin endometriosis and adenomyosis. Ann. N. \nY. Acad. Sci. 2007; 1101: 38–48. doi:10.1196/\nannals.1389 .036\nKuijsters, N.P .M., Methorst, W.G., Kortenhorst, \nM.S.Q., Rabotti, C., Mischi, M., Schoot, B.C. \nUterine peristalsis and fertility: current \nknowledge and future perspectives: a review \nand meta-analysis. Reprod. Biomed. Online \n2017; 35: 50–71. doi:10.1016/J.RBMO.2017 .03.019\nKuijsters, N.P .M., Sammali, F., Rabotti, C., \nHuang, Y., Mischi, M., Schoot, B.C. Visual \ninspection of transvaginal ultrasound videos \nto characterize uterine peristalsis: an inter-\nobserver agreement study. J. Ultrasound \n2020; 23: 37–44. doi:10.1007/s40477-018-\n00356-z\nLeyendecker, G., Kunz, G., Wildt, L., Beil, \nD., Deininger, H. Uterine hyperperistalsis \nand dysperistalsis as dysfunctions of the \nmechanism of rapid sperm transport in \npatients with endometriosis and infertility. \nHum. Reprod. 1996; 11: 1542–1551. doi:10.1093/\noxfordjournals.humrep.a019435\n\n122  RBMO  VOLUME 46  ISSUE 1  2023\nSammali, F., Blank, C., Bakkes, T.H.G.F., Huang, \nY., Rabotti, C., Schoot, B.C., Mischi, M.  \nPrediction of embryo implantation by \nmachine learning based on ultrasound strain \nimaging.  IEEE International Ultrasonics \nSymposium, IUS 2019: 1141–1144. doi: 10.1109/\nULTSYM.2019 .8926228\nSammali, F., Blank, C., Bakkes, T.G.H., \nHuang, Y., Rabotti, C., Schoot, B.C., \nMischi, M. Multi-modal uterine-activity \nmeasurements for prediction of embryo \nimplantation by machine learning. IEEE \nAccess 2021; 9: 47096–47111. doi:10.1109/\nACCESS.2021.3067716\nSammali, F., Blank, C., Huang, Y., Kuijsters, \nN.P .M., Rabotti, C., Schoot, B.C., Mischi, \nM. Quantitative analysis of uterine motion \noutside pregnancy by dedicated ultrasound \nspeckle tracking.  IEEE International \nUltrasonics Symposium, IUS 2018: 1–4. \ndoi:10.1109/ULTSYM.2018.8579772\nSammali, F., Blank, C., Xu, L., Huang, Y., \nKuijsters, N.P .M., Schoot, B.C., Mischi, M.  \nExperimental setup for objective evaluation \nof uterine motion analysis by ultrasound \nspeckle tracking.  Biomed. Phys. Eng. Express \n2018; 4035012. doi: 10.1088/2057-1976/\naab053\nSammali, F., Kuijsters, N.P .M., Huang, Y., Blank, \nC., Rabotti, C., Schoot, B.C., Mischi, M. \nDedicated ultrasound speckle tracking for \nquantitative analysis of uterine motion outside \npregnancy. IEEE Trans. Ultrason. Ferroelectr. \nFreq. Control. 2019; 66: 581–590. doi:10.1109/\nTUFFC.2018.2867098\nVan den Bosch, T., de Bruijn, A.M., de Leeuw, \nR.A., Dueholm, M., Exacoustos, C., Valentin, \nL., Bourne, T., Timmerman, D., Huirne, J.A.F. \nSonographic classification and reporting \nsystem for diagnosing adenomyosis. \nUltrasound Obstet. Gynecol. 2019; 53: \n576–582. doi:10.1002/uog.19096\nvan Gestel, I., Ijland, M.M., Hoogland, H.J., Evers, \nJ.L.H. Endometrial wave-like activity in the \nnon-pregnant uterus. Hum. Reprod. Update \n2003; 9: 131–138. doi:10.1093/humupd/dmg011\nVon Elm, E., Altman, D.G., Egger, M., Pocock, \nS.J., Gøtzsche, P .C., Vandenbroucke, J.P . The \nStrengthening the Reporting of Observational \nStudies in Epidemiology (STROBE) statement: \nguidelines for reporting observational studies. \nBull. World Health Organ. 2007; 85: 867–872. \ndoi:10.2471/BLT.07 .045120\nReceived 31 May 2022; received in revised form \n16 August 2022; accepted 18 August 2022.","source_license":"CC0","license_restricted":false}